[From Bill Powers (940526.0910 MDT)]
Bill Cunningham (940526.0800) --
I'm sure we do sometimes control for reducing uncertainty itself.
But this is the very specific circumstance in which it is not-being-
sure itself that is the error, like not knowing what to say to a
dinner companion or a Queen. We always control for specific
perceptions relative to specific reference levels. Most of our
perceptions are not concerned with certainty or uncertainty, but
with things like how hard we are pushing, how much salt is in the
soup, whether we are singing on pitch, and whether the car is in its
lane. In a formal sense we can always associate some uncertainty
with those other perceptions, but that means only that there is some
noise level; it doesn't mean that the object of action is to bring
perceived uncertainty to any particular state. If I am controlling
the position of a car, that is the perception I am controlling, not
my uncertainty about the position of the car -- even though it can
be shown that there is in fact some uncertainty, some random
variation, in that perception, relative to an external measure of
the position of the car.
I have no problem in agreeing that controlling for uncertainty can
be done and is done. Sometimes, as in games, we even try to control
for maximum uncertainty -- on the part of an opponent. But I think
that uncertainty is just another kind of perceptual variable, among
many.
ยทยทยท
---------------------------------------------------------------
Best,
Bill P.